Changeset 11008 for trunk/gsdl

Timestamp:

2005-12-19T15:17:49+13:00 (18 years ago)

Author:

mdewsnip

Message:

Added an option to run the "fribidi" Unicode Bidirectional Algorithm program over the converted file, primarily to fix up right-to-left PDFs that pdftohtml can't handle.

Location:

trunk/gsdl/perllib

Files:

• plugins/ConvertToPlug.pm (modified) (2 diffs)
• strings.rb (modified) (1 diff)

trunk/gsdl/perllib/plugins/ConvertToPlug.pm

@@ -67 +67 @@
     #'type' => "regexp",
     'deft' => "" },
+      { 'name' => "apply_fribidi",
+    'desc' => "{ConvertToPlug.apply_fribidi}",
+    'type' => "flag",
+    'reqd' => "no" },
       { 'name' => "use_strings",
     'desc' => "{ConvertToPlug.use_strings}",
@@ -345 +349 @@
     $self->{'conv_filename'} = $conv_filename;
     $self->convert_post_process($conv_filename);
-    
+
+    # Run the "fribidi" (http://fribidi.org) Unicode Bidirectional Algorithm program over the converted file
+    # Added for fixing up Persian PDFs after being processed by pdftohtml, but may be useful in other cases too
+    if ($self->{'apply_fribidi'} && $self->{'converted_to'} =~ /(HTML|TEXT)/) {
+    my $fribidi_command = "fribidi \"$conv_filename\" >\"${conv_filename}.tmp\"";
+    if (system($fribidi_command) != 0) {
+        print STDERR "ERROR: Cannot run fribidi on \"$conv_filename\".\n";
+    }
+    else {
+        &util::mv("${conv_filename}.tmp", $conv_filename);
+    }   
+    }
+
     my $secondary_plugins =  $self->{'secondary_plugins'};
     my $num_secondary_plugins = scalar(keys %$secondary_plugins);

trunk/gsdl/perllib/strings.rb

@@ -621 +621 @@
 BRSPlug.desc: BRSPlug deals with the BBC collections. There are three BBC collections we deliver, which are National Sound Archive (nsa), British Film Institute (bfi), and BBC Other (bbcother) (everything does not fit into one of the first two). Four options are provided which are all-build all collections, nsa-build the NSA collections, bfi-build the BFI collections, bbcother-build the OTHER collections.
 
+ConvertToPlug.apply_fribidi:Run the "fribidi" Unicode Bidirectional Algorithm program over the converted file (for right-to-left text).
 ConvertToPlug.convert_to:Plugin converts to TEXT or HTML or various types of Image (e.g. JPEG, GIF, PNG).
 ConvertToPlug.convert_to.auto:Automatically select the format converted too.  Format chosen depends on input document type, for example Word will automatically be converted to HTML, whereas PowerPoint will be converted to Greenstone's PagedImage format.